ETX Sightline - The 2025 Review

Nine months in: what ETX Upstream built in 2025

2025 will stand out as a wild year for a lot of reasons. The AI buildout went from a forecast to trillions of dollars of committed capex. Interconnection queues backed up further than they ever have. And the physical constraints on where and how fast power actually gets built stopped being a footnote to the AI story and became the story — the binding constraint everyone from hyperscalers to utilities to their own investors started asking about directly.

That's the year ETX Upstream launched into. We're writing this at the nine-month mark, which by our own count is not very long at all. But across the three areas we organize the business around — Land, Load, and Enablement — it's already been a full year's worth of work.

1) Land

We originated and now manage 3,292 acres across our energy, data, and carbon basin strategies. That number undersells what it represents, because raw acreage was never really the point — it's a big part of why we're as obsessive as we are about what a parcel actually nets out to once every real-world constraint is accounted for, not just what the deed says.

What's sitting on that footprint by year-end is the more interesting figure: operative hosting capacity for 15+ MW of solar, over 1+ GW of large load development, 500+ MW of thermal generation development, 100,000 square feet of industrial facilities, and 9,500 tons CO2e/year of carbon sequestration capacity.

Five different end uses, one underlying constrained asset class. Land is the upstream input every one of those categories is competing for.

2) Load

Two flagship, gigawatt-scale Energy Hub projects are now under development, both grown out of sustained, close engagement with public power utilities in the Southeast.

Utilities in that seat are threading a genuinely hard needle right now: absorbing enormous new demand without destabilizing service to everyone already on the system, on a timeline that used to be measured in years and is increasingly measured in quarters.

We also signed an MOU with a leading independent power producer to co-locate and co-develop a combined generation-and-load hybrid. This is a structural fix for the interconnection-queue problem the whole industry is living through: instead of a load project waiting years behind unrelated generation in the same queue, pairing generation and load from day one collapses that timeline. It's a small change to deal structure with outsized consequences for how fast power actually reaches the ground.

3) Enablement

The most consequential work this year happened inside the platform itself.

LandCommand was reengineered — version 3.0 shipped with a genuinely dynamic architecture and materially expanded AI capabilities. This is the identity-layer work underneath everything else on this list: every acre in the Land portfolio and every megawatt in the Load pipeline has been found, scored, and de-risked by LandCommand — software that can reason over physical space the way markets already reason over securities.

Separately, another 620 MWdc of solar generation reached Notice to Proceed this year through our legacy Galehead development pipeline. NTP is the moment a project stops being a plan and starts being construction — permits in hand, financing committed, steel in the ground next. That's 620 MWdc of real generation that started as line algorithms in LandCommand and is now being built.

Land, Load, and Enablement aren't three separate initiatives so much as three views of the same underlying bet: that the physical world — the acreage, the megawatts, the interconnection queues — is going to keep being the actual constraint on how fast this industry can move, and that the firms who can see and act on that constraint fastest win.

2025 was the year we built the base layer for doing that at scale. 2026 is the year we find out how far it goes.

Wishing the same for everyone in our network. Here's to a bold 2026!